Models

Discover and compare available AI models, their capabilities, rankings, and benchmark scores.

Endpoint

GET https://conduit.im/api/rankings

This endpoint is public and does not require authentication. Responses are cached for up to 1 hour.

Query Parameters

Parameter	Type	Default	Description
source	string	`INTERNAL`	Ranking source. One of: `INTERNAL`, `LMSYS_ARENA`, `HUGGINGFACE`, `ARTIFICIAL_ANALYSIS`, `OPENROUTER`
category	string	—	Filter by model category. One of: `GENERAL`, `CODING`, `MATH`, `REASONING`, `CREATIVE`, `OPEN_SOURCE`
limit	integer	`50`	Number of results to return (1–100)
offset	integer	`0`	Number of results to skip for pagination
sortBy	string	`rank`	Sort field. One of: `rank`, `score`, `name`, `change`
sortOrder	string	`asc`	Sort direction. One of: `asc`, `desc`
search	string	—	Search by model name, provider, or provider ID (case-insensitive)
commercialOnly	boolean	—	When `true`, only return models available for commercial use

Example Request

Fetch the top 10 models in the CODING category, sorted by rank:

curl "https://conduit.im/api/rankings?category=CODING&limit=10&sortBy=rank&sortOrder=asc"

Response Format

The response contains ranked model data, pagination info, source health status, and a timestamp.

{
  "rankings": [
    {
      "rank": 1,
      "categoryRank": 1,
      "change": 0,
      "model": {
        "id": "claude-opus-4-6",
        "name": "Claude Opus 4.6",
        "provider": "Anthropic",
        "category": "CODING",
        "contextWindow": 200000,
        "imageSupport": true,
        "functionCalling": true,
        "commercialUse": true,
        "license": "Commercial"
      },
      "scores": {
        "overall": 94.2,
        "quality": 96.1,
        "speed": 78.5,
        "cost": 62.0,
        "elo": 1285,
        "benchmarks": {
          "mmlu": 92.3,
          "hellaswag": 95.1,
          "humanEval": 90.8
        }
      },
      "metadata": {
        "confidence": 0.95,
        "dataFreshnessHours": 2,
        "lastUpdated": "2026-03-09T12:00:00.000Z"
      }
    },
    {
      "rank": 2,
      "categoryRank": 2,
      "change": 1,
      "model": {
        "id": "gpt-4o",
        "name": "GPT-4o",
        "provider": "OpenAI",
        "category": "CODING",
        "contextWindow": 128000,
        "imageSupport": true,
        "functionCalling": true,
        "commercialUse": true,
        "license": "Commercial"
      },
      "scores": {
        "overall": 91.8,
        "quality": 93.4,
        "speed": 85.2,
        "cost": 70.5,
        "elo": 1260,
        "benchmarks": {
          "mmlu": 90.1,
          "hellaswag": 93.7,
          "humanEval": 88.4
        }
      },
      "metadata": {
        "confidence": 0.93,
        "dataFreshnessHours": 3,
        "lastUpdated": "2026-03-09T11:30:00.000Z"
      }
    }
  ],
  "pagination": {
    "total": 42,
    "limit": 10,
    "offset": 0,
    "hasMore": true
  },
  "sources": [
    {
      "source": "INTERNAL",
      "isHealthy": true,
      "lastSyncAt": "2026-03-09T10:00:00.000Z"
    }
  ],
  "timestamp": "2026-03-09T12:05:00.000Z"
}

Model Properties

Field	Type	Description
id	string	Provider model ID (e.g., `gpt-4o`, `claude-opus-4-6`). Use this as the `model` parameter in Chat Completions.
name	string	Human-readable model name
provider	string	Model provider (e.g., OpenAI, Anthropic, Google)
category	string	Model category: `GENERAL`, `CODING`, `MATH`, `REASONING`, `CREATIVE`, or `OPEN_SOURCE`
contextWindow	integer	Maximum context window size in tokens
imageSupport	boolean	Whether the model supports image inputs
functionCalling	boolean	Whether the model supports function/tool calling
commercialUse	boolean	Whether the model is available for commercial use
license	string \| null	Model license type (e.g., "Commercial", "Apache 2.0", "MIT")

Scores & Benchmarks

Each ranked model includes composite scores and individual benchmark results. All scores range from 0 to 100 unless otherwise noted.

Score	Range	Description
overall	0–100	Weighted composite score across all dimensions
quality	0–100	Output quality and accuracy score
speed	0–100	Inference speed and latency score
cost	0–100	Cost efficiency score (higher = more cost-effective)
elo	~800–1400	Elo rating from head-to-head model comparisons
benchmarks.mmlu	0–100	MMLU (Massive Multitask Language Understanding) score
benchmarks.hellaswag	0–100	HellaSwag commonsense reasoning score
benchmarks.humanEval	0–100	HumanEval code generation score

Filtering & Searching

Combine query parameters to narrow results. Here are some common patterns:

Search by name or provider:

curl "https://conduit.im/api/rankings?search=claude"

Top 5 commercially-usable reasoning models:

curl "https://conduit.im/api/rankings?category=REASONING&commercialOnly=true&limit=5"

Models sorted by score (highest first):

curl "https://conduit.im/api/rankings?sortBy=score&sortOrder=desc&limit=20"

Paginate through results:

# Page 1
curl "https://conduit.im/api/rankings?limit=10&offset=0"

# Page 2
curl "https://conduit.im/api/rankings?limit=10&offset=10"

Using Models with Chat Completions

The model.id field in the rankings response (e.g., gpt-4o, claude-opus-4-6) is the value you pass as the model parameter to the Chat Completions API.

curl -X POST "https://api.conduit.im/v1/chat/completions" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-opus-4-6",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

Next Steps

Chat Completions API — Send requests to the models you discover here
Error Handling — Handle API errors gracefully
Live Rankings — Explore model rankings interactively

Documentation